« Changelog History


Redash v9.0.0-beta Release Notes

Release Date: 2020-06-11 // almost 4 years ago

  • 🚀 This release was long time in the making and has several major changes:

    • ⚡️ Our backend code was updated to support Python 3 and we no longer support Python 2. If you're using our Docker images, this should be a transparent change for you.
    • ⚡️ We replaced Celery with RQ for background jobs processing. This will require some setup updates -- see instructions below.
    • 🚚 The frontend code is now 100% React and we removed all the Angular dependencies.

    🚀 This release was made possible by contributions from over 50 people: @ari-e, @ariarijp, @arihantsurana, @arikfr, @atharvai, @cemremengu, @chulucninh09, @citrin, @daniellangnet, @DavidHernandez, @deecay, @dmudro, @erans, @erels, @ezkl, @gabrieldutra, @gstaykov, @ialeinikov, @ikenji, @Jakdaw, @jezdez, @juanvasquezreyes, @koooge, @kravets-levko, @kykrueger, @leibowitz, @leosunmo, @lihan, @loganprice, @mickeey2525, @mnoorenberghe, @monicagangwar, @NicolasLM, @p-yang, @Ralnoc, @ranbena, @randyzwitch, @rauchy, @rxin, @saravananselvamohan, @satyamkrishna, @shinsuke-nara, @stefan-mees, @stevebuckingham, @susodapop, @taminif, @thewarpaint, @tsuyoshizawa, @uncletimmy3, @wengkham.

    ⬆️ Upgrading

    ⬆️ Typically, if you are running your own instance of Redash and wish to upgrade, you would simply modify the Docker tag in your docker-compose.yml file. Since RQ has replaced Celery in this version, there are a couple extra modifications that need to be done in your docker-compose.yml:

    ⏱ 1. Under services/scheduler/environment, omit QUEUES and WORKERS_COUNT (and omit environment altogether if it is empty). 👷 2. Under services, add a new service for general RQ jobs:

    worker:
  <<: *redash-service
  command: worker
  environment:
    QUEUES: "periodic emails default"
    WORKERS_COUNT: 1

    🏗 Following that, force a recreation of your containers with docker-compose up --force-recreate --build and you should be good to go.

    UX

    • Redesigned Query Results page:
      • Completely new layout is easier to read for non-technical Redash users.
      • Empty query results are clearly displayed. User is now prompted to edit or execute the query.
    • Mobile Experience Improvements:
      • UI element spacing has been redesigned for clarity
      • Admin pages now honor max-width. Tables scroll independent of the top menu.
      • Large legends no longer shrink the visualization on small screens.
      • Fix: it was sometimes impossible to scroll pages with dashboards because the visualizations captured every touch event.
      • Fix: Visualizations on small screens would not always show horizontal scroll bars.
    • Dashboards can now be un-archived using the API.
    • 🐎 Dashboard UI performance was improved.
    • List pages were changed to show a user's name instead of avatar.
    • Search-enabled tables now show a prompt for which columns will be searched.
    • In the visualization editor, the settings pane now scrolls independent of the visualization preview.
    • Tokens in the schema viewer now sort alphabetically.
    • 🔗 Links to settings panes that require Admin privileges are now hidden from non-Admins.
    • The Admin page now remembers which tab you were viewing after a page reload.

    Visualizations

    • 🔋 Feature: Allow bubble size control with either coefficient or sizemode.
    • 🔋 Feature: Table visualization now treats Unix timestamps in query results as timestamps.
    • 🔋 Feature: It's now possible to provide a description to each Table column, appearing in UI as a tooltip.
    • 🔋 Feature: Added tooltip and popover templating to the map with markers visualization.
    • 🔋 Feature: Added an organization setting to hide the Plotly mode bar on all visualizations.
    • 🔋 Feature: Cohort visualization now has appearance settings.
    • 🔋 Feature: Add option to explicitly set Chart legend position.
    • 🔄 Change: Deprecated visualizations are now hidden.
    • 🔄 Change: Table settings editor now extends vertically instead of horizontally.
    • 🔄 Change: The maximum table pagination is now 500.
    • 🔄 Change: Pie chart labels maintain contrast against lighter slices.
    • 🛠 Fix: Chart series switched places when picking Y axis.
    • 🛠 Fix: Third column was not selectable for Bubble and Heatmap charts.
    • 🛠 Fix: On the counter visualizations, the “count rows” option showed an empty string instead of 0.
    • 🛠 Fix: Table visualization with column named "children" rendered +/- buttons.
    • 🛠 Fix: Sankey visualization now correctly occupies all available area even with fewer stages.
    • 🛠 Fix: Pie chart ignores series labels.

    Data Sources

    • 🆕 New Data Sources: Amazon Cloudwatch, Amazon CloudWatch Logs Insights, Azure Kusto, Exasol.
    • Athena:
      • Added the option to specify a base cost in settings, displaying a price for each query when executed.
    • BigQuery:
      • Fix: large jobs continued running after the user clicked “Cancel” query execution.
    • Cassandra:
      • Updated driver to 3.21.0 which dramatically reduces Docker build times.
      • SSL options are now available.
    • Clickhouse:
      • You can now choose whether to verify the SSL certificate.
    • Databricks:
      • Databricks now use an ODBC-based connector.
      • Fix: Date column was coerced to DateTime in the front-end.
    • Druid:
      • Added username and password authentication option.
    • Microsoft SQL Server
      • Added support for ODBC connections via pyodbc. There are now two MSSQL data source types. One using TDS. The other is using ODBC.
    • MongoDB:
      • Added support for running queries on secondary in replicaset mode.
      • Fix: Connection test always succeeded.
    • Oracle:
      • Fix: Connection would fail if username or password contained special characters.
      • Fix: Comparisons would fail if scale was None.
    • RDS:
      • Updated rds-combined-ca-bundle.pem to the latest CA.
    • Redshift:
      • Added the ability to use IAM Roles and Users.
      • Fix: Redshift was unable to have its schema refreshed.
    • Rockset:
      • Fix: Allow Redash to load collections in all workspaces.
    • Snowflake:
      • You can now refresh the snowflake schema without waking the cluster.
      • Added support for all of Snowflake’s datetime types. Otherwise certain timestamps would only appear as strings in the front-end.
    • TreasureData:
      • Fix: API calls would fail when setting a non-default region.

    Alerts

    • 🔋 Feature: Added ability to mute alerts without deleting them.
    • 🛠 Fix: numerical comparisons failed if value from query was a string.

    Parameters

    • ➕ Added Last x Days options for date range parameters.
    • 🛠 Fix: Parameters added in empty queries were always added as text parameters

    🐛 Bug Fixes

    • 🛠 Fix: Alembic migration schema was preventing v4 users from upgrading. In v5 we started encrypting data source credentials in the database.
    • 🛠 Fix: System admin dashboard would not show correct database size if non-default name was used.
    • 🛠 Fix: refresh_queries job would break if any query had a bad schedule object.
    • 🛠 Fix: Orgs with LDAP enabled couldn’t disable password login.
    • 🛠 Fix: SSL mode was sometimes sent as an empty string to the database instead of omitted entirely.
    • 🛠 Fix: When creating new Map visualization with clustering disabled, map would crash on save.
    • 🛠 Fix: It was possible on the New Query page to click “Save” multiple times, causing multiple new query records to be created.
    • 🛠 Fix: Visualization render errors on a dashboard would crash the entire page.
    • 🛠 Fix: A scheduled execution failure would modify the query’s “updated_at” timestamp.
    • 🛠 Fix: Parameter UI would wrap awkwardly during some drag operations.
    • 🛠 Fix: In dashboard edit mode, users couldn’t modify widgets.
    • 🛠 Fix: Frontend error when parsing a NaN float.

    Other

    • ➕ Added TSV as a download format (in addition to CSV and Excel).
    • ➕ Added maildev settings (helps with automated settings).
    • Refine permissions usage in Redash to allow for guest users
    • The query results API now explicitly handles 404 errors.
    • Forked queries now retain the tags of the original query.
    • We now allow setting custom Sentry environments.
    • 👕 Started using Black linter for our Python source code
    • ➕ Added CLI command to re-encrypt data source details with new secret key.
    • Favorites list is now loaded on menu click instead of on page load.
    • Administrators can now allow connections to private IP addresses.